Description
Data analysis and integration are the main objectives of any database based project for a particular organization, and there are several ways described by field experts to achieve this objective. Amongst various approaches to data base management, Data Warehousing strategies are preferred over other methods when it comes to business data. Herein, data related to specific departments or entities is extracted from operational system, processed, put into the specialized relational databases, organized and presented as actionable information. A Data Warehouse is most often based on a decisional-architecture platform, which can be used by analytical and reporting tools, serve as a digital dashboard to company executives, and provide analysis with easy-to-access data irrespective of source system. It also improves data accuracy and keeps a track of historical data. A Data Mart is essentially a subset of a Data Warehouse. It is a flexible set of data, ideally based on the most atomic (granular) data possible to extract from an operational source, presented in a symmetric (dimensional) model, and most resilient when faced with unexpected user queries. In its most simplistic form, a Data Mart represents data from a single business process or function. The process, from its design till commissioning, entails low initial investment. The aim of this thesis work is to design a sales-function Data Mart, using star schema strategy. It involves collecting requirements from the sales department of a given organization, and discussing a flexible Data Mart architecture to implement those requirements and to facilitate its seamless integration with Data Marts representing other business processes of the same organization. The thesis builds the prototype of the sales data mart as applicable to a cellular telephone manufacturing company and discusses the overall architecture at an enterprise level data warehousing, including business requirements, data mart architecture, and logical and physical model design. Further, this thesis follows an "incremental approach" for building the data mart.
